Tell HN: Anthropic no longer allowing Claude Code subscriptions to use OpenClaw
Claude Code users must now pay extra for third-party tools like OpenClaw starting April 4.
Anthropic has announced a significant policy change affecting users of its Claude Code subscription service. Starting April 4 at 12pm PT, subscribers will no longer be able to use their subscription tokens with third-party harnesses like OpenClaw. Instead, these tools will require separate pay-as-you-go billing through an 'extra usage' system billed independently from subscriptions. The company is offering affected users a one-time credit equal to their monthly subscription price, redeemable by April 17, along with discounts up to 30% for pre-purchasing usage bundles.
In an email to users, Anthropic explained that third-party harnesses 'put an outsized strain on our systems' and that 'capacity is a resource we manage carefully.' The company stated it needs to 'prioritize our customers using our core products.' While OpenClaw is the first tool affected, the policy applies to all third-party harnesses and will be rolled out to more shortly. Users who prefer can request a subscription refund, with details coming in a follow-up email.
The move has sparked discussion about subscription economics and autonomous AI usage patterns. Some commenters note that tools like OpenClaw can consume 6-8 times more resources than typical human users, creating cost pressures. Others suggest this is less about capacity management and more about keeping users within Anthropic's ecosystem, as similar high-usage patterns can be achieved with native tools like Claude's /loop feature and scheduled tasks.
- OpenClaw and all third-party harnesses now require separate pay-as-you-go billing starting April 4
- Anthropic cites 'outsized strain on our systems' and offers one-time credit equal to monthly subscription
- Policy affects all Claude Code subscriptions, with refund option available for dissatisfied users
Why It Matters
Shifts AI subscription economics, impacting developers using autonomous agents and signaling tighter platform control.